This is a n Op e n Acces s doc u m e n t dow nloa d e d fro m ORCA, Ca r diff U nive r si ty 'sins ti t u tion al r e posi to ry: h t t p s://o rc a.c a r diff.ac.uk/124 5 9 2/

This is t h e a u t ho r’s ve r sion of a wo rk t h a t w as s u b mi t t e d to / a c c e p t e d forp u blica tion.

Cit a tion for final p u blish e d ve r sion:

M a cino, M a r g h e ri t a , Ba r n e s , Alexa n d r a J., Alth a h b a n, S ul t a n M., Qu, Ruiya n g,Gibson, E m m a K., Mo r g a n, David J, F r e a kley, Simo n J., Dimit r a to s, Nikolaos,

Kiely, Ch ris top h e r J., Gao, Xiang, Be ale, Andr e w M., Be t h ell, Don ald, H e, Qia n,M e e n a ks his u n d a r a m, S a nk a r a n d H u t c hin gs, Gr a h a m J. 2 0 1 9. Tuning of

c a t alytic si t e s in P t/TiO2 c a t alys t s for c h e m os elec tive hyd rog e n a tion of 3-ni t ro s ty r e n e . N a t u r e Ca t alysis 2 , p p . 8 7 3-8 8 1. 1 0.10 3 8/s41 9 2 9-0 1 9-0 3 3 4-3

file

P u blish e r s p a g e: h t t p s://doi.or g/10.1 03 8/s 4 1 9 2 9-0 1 9-0 3 3 4-3< h t t p s://doi.o rg/10.10 3 8/s4 1 9 2 9-0 1 9-0 3 3 4-3 >

Ple a s e no t e: Ch a n g e s m a d e a s a r e s ul t of p u blishing p roc e s s e s s uc h a s copy-e di ting,

for m a t ting a n d p a g e n u m b e r s m ay no t b e r eflec t e d in t his ve r sion. For t h ed efini tive ve r sion of t his p u blica tion, ple a s e r ef e r to t h e p u blish e d sou rc e. You

a r e a dvise d to cons ul t t h e p u blish e r’s ve r sion if you wish to ci t e t his p a p er.

This ve r sion is b ein g m a d e av ailable in a cco r d a n c e wit h p u blish e r policie s.S e e

h t t p://o rc a .cf.ac.uk/policies.h t ml for u s a g e policies. Copyrigh t a n d m o r al r i gh t sfor p u blica tions m a d e available in ORCA a r e r e t ain e d by t h e copyrig h t

hold e r s .

Tuning of catalytic sites in Pt/TiO2 catalysts for the chemoselective

hydrogenation of 3-nitrostyrene

Margherita Macino,a Alexandra J. Barnes,a Sultan M. Althahban,b Ruiyang Qu,c,d

Emma Gibsone, David J Morgana, Simon J. Freakleya, Nikolaos Dimitratosa,f,

Christopher J. Kiely,a,b Xiang Gao,c,d Andrew M. Beale,g,h Donald Bethell,i Qian He,a

Meenakshisundaram Sankar,a* and Graham J. Hutchingsa*

a Cardiff Catalysis Institute, Cardiff University, Main Building, Park Place, Cardiff

CF10.3AT, UK.

b Department of Materials Science and Engineering, Lehigh University, 5 East Packer

Avenue, Bethlehem, PA 18015-3195, USA.

c State Key Laboratory of Clean Energy Utilization, College of Energy Engineering,

Zhejiang University, Hangzhou, 310027, China.

d Institute for Thermal Power Engineering, Zhejiang University, Hangzhou, 310027,

P. R. China.

e School of Chemistry, Joseph Black Building, University Avenue, Glasgow, G12

8QQ, UK

f Department of Industrial Chemistry "Toso Montanari", Alma Mater Studiorum-

University of Bologna, 40136 Bologna, Italy.

g UK Catalysis Hub, Research Complex at Harwell (RCaH), Rutherford Appleton

Laboratory, Harwell Oxon, OX11 0FA, UK.

hDepartment of Chemistry, University College London, 20 Gordon Street, London,

WC1H 0AJ, UK.

i Department of Chemistry, University of Liverpool, Liverpool L69 3BX, UK.

Corresponding Authors: Dr. M. Sankar, Cardiff University, Email:

Sankar@cardiff.ac.uk ; Prof. G. J. Hutchings, Cardiff University, Email:

Hutch@cardiff.ac.uk.

Abstract

Tuning the catalytic activities of supported metal nanoparticles can be achieved by

manipulating their structural properties using an appropriate design and synthesis

strategy. Each step in a catalyst synthesis method can potentially play an important

role in preparing the most efficient catalyst for a particular chemical reaction. Here

we report the careful manipulation of the post-synthetic heat treatment procedure,

together with control over the the amount of metal loading, to prepare a highly

efficient 0.2 wt.% Pt/TiO2 catalyst for the chemoselective hydrogenation of 3-

nitrostyrene. We found that for Pt/TiO2 catalysts with 0.2 and 0.5wt.% loading levels,

reduction alone at 450 °C induces the coverage of Pt nanoparticles by TiOx through a

strong metal support interaction which is detrimental for their catalytic activities.

However, this surface coverage can be avoided by combining a calcination treatment

at 450 °C with a subsequent reduction treatment at 450 °C allowing us to prepare a

exceptionally active Pt/TiO2 catalyst with the optimum Pt distribution. Detailed

characterisation of these catalysts has revealed that the peripheral sites at the Pt

metal/TiO2 support interface are the most likely active sites for this hydrogenation

reaction.

Introduction

Supported metal nanocatalysts play a crucial role in our endeavour to develop new

technologies for energy and environmental applications, including the production of

fine chemicals and fuels from renewable and sustainable feedstock while at the same

time achieving reduced emission targets.1,2 Selective hydrogenation reactions,

catalysed by supported metal nanoparticles, are one of the most widely studied

reactions in academia and industry.3 One typical example is the hydrogenation of

nitroarenes to produce functionalised anilines, which are commonly used in the

production of a number of fine and bulk chemicals including polymers, fertilisers,

pharmaceuticals, dyes and biologically active compounds.4,5 Recently researchers

have focused on the chemoselective hydrogenation of nitro aromatics to form

anilines, especially in the presence of other reducible groups such as C=C, C=O, C≡C

and C≡N, because of the high synthetic value of these compounds6,7 Many reduction

processes in industry still make use of stoichiometric reducing agents like sodium

hydrosulphite (or H2S), iron in acidic media (Bechamp process), stannous chloride,

samarium iodide or zinc in ammonium hydroxide.8-10 These methods are not only

unselective, but also generate huge amounts of unwanted waste. Catalytic reduction,

using molecular H2, is highly desirable because of its economics and availability; it is

also environmentally benign since water is the only by-product produced.3,7,11,12

However, achieving high chemoselectivity in this reaction without adding any

homogeneous additives is still challenging. When hydrogen donors such as

hydrazine,13,14 formic acid15 or sodium borohydride16,17 are used, high

chemoselectivities for nitro group reduction can be achieved. Recent advances in

controlling the structure and properties of supported metal nanoparticles have created

a renewed interest in the development of better catalysts for chemoselective

hydrogenation reactions using molecular H2.3 For Pt/TiO2 catalysts, alloying with a

second metal (e.g., Au, Ni or Zn) has also been reported to increase the

chemoselectivity.11 Corma et al. showed that monometallic Au nanoparticles (< 5nm)

supported on Fe2O3 could increase the selectivity (up to 99%) towards nitro group

reduction during the hydrogenation of nitrostyrene.18 Zhang et al. reported that atomic

dispersions of Pt on FeOx were very active catalysts for the chemoselective

hydrogenation of nitrostyrene19 and furthermore they found that adding alkali metals

to high loading Pt/FeOx catalysts, where nanoparticles are also present, increased the

chemoselectivity dramatically.20

Scheme-1: Schematic representation of the chemoselective hydrogenation of 3-nitrostyrene

(3-NS) to form the desired product 3-vinylaniline (3-VA) and associated side products 3-

ethylnitrobenzene (3-ENB), and 3-ethylaniline (3-EA).

Every step in the catalyst synthesis process plays an important role in tuning the

structural characteristics of the catalyst and in turn its activity, selectivity and stability.

We have reported that addition of excess of chloride ions during wet-impregnation

helps in controlling the particle size and composition of supported AuPd

nanoparticles.21 Similarly, heat treatment procedures are critical in controlling metal

particle size, morphology,22 alloy formation,21 and for removal of organic ligands or

poisons, 23 thereby dramatically affecting the catalytic performance.21,24,25 Typical

heat treatments employed for hydrogenation catalysts are as follows: the as-

synthesized samples are usually calcined first, to decompose the metal precursors on

the support surface, followed by a gas-phase reduction, either in-situ or ex-situ, to

reduce the metal ions to form the metal nanoparticles.11,19 Corma et al. controlled the

particle size and morphology of Pt nanoparticles supported on TiO2, by choosing an

appropriate activation temperature to tune the chemoselectivity during the

hydrogenation of nitrostyrene.11 Studies on the inter-relationship between metal

loading and heat treatment are highly important, because of the current drive to

reduce the metal loading to efficiently utilise scarce and expensive noble metals.19

Here we report the catalytic performance of 0.05 wt.%, 0.08 wt.%, 0.2 wt.% and 0.5

wt.% Pt on TiO2 for the chemoselective hydrogenation of 3-nitrostyrene (3-NS) to 3-

vinylaniline (3-VA) (Scheme-1). This substrate was selected as there are two

potentially reducible groups making it a widely used model substrate for

chemoselective hydrogenation reactions. The active sites have been fine-tuned by

optimising the post synthesis heat treatment protocol by systematically applying

different high temperature heat treatments (i.e., either reduction only or calcination

only or calcination followed by reduction) to all these catalysts and we report the high

catalytic activity displayed by the 0.2 wt.% Pt/TiO2 calc+red sample. All these

catalyst variants were characterised by X-ray photoelectron spectroscopy (XPS),

scanning transmission electron microscopy (STEM), extended X-ray absorption fine

structure (EXAFS) and X-ray absorption near edge structure (XANES) as well as CO

DRIFTS. This has allowed us to explore the complex interplay between Pt loading

level, heat treatment procedure and strong metal support interactions (SMSI). We

deduce that the extent of formation of exposed Pt peripheral sites with the TiO2

support directly correlates with the activity of these Pt/TiO2 catalysts.

Results

Hydrogenation of 3-nitrostyrene

Initially, a 0.2 wt.% Pt/TiO2 catalyst reduced at 450 °C, was tested for the liquid

phase hydrogenation of 3-nitrostyrene (3-NS) at 40 °C under 3 bar H2 pressure with

toluene as the solvent. 3-NS was rapidly and selectively converted into 3-VA; plots of

3-NS concentration versus time were linear up to about 25% conversion (SI, Fig. S1),

giving an initial conversion rate of 1.55x10-3 mol3NS L-1 min-1 with greater than 99%

selectivity towards 3-vinylaniline (3-VA) (Figure 1(a)). Initial rates with different

starting concentrations of 3-NS were used to confirm that reactions are zero order in

3-NS (SI, Figs. S2 & S3) for both 0.2 wt.% Pt/TiO2 and 0.08 wt.% Pt/TiO2. Other

potential products such as (i) 3-ethyl nitrobenzene (3-ENB) formed by the selective

hydrogenation of C=C and (ii) 3-ethylaniline (3-EA) formed by the complete

hydrogenation of both the nitro and C=C groups were either not detected or were

present only in trace quantities. In line with many other reports, the fully

hydrogenated 3-ethyl-1-cyclohexylamine product was not detected under our reaction

conditions; hence it is not included in the discussion (Scheme-1). Inspired by the

work of Corma et al.11, one portion of the dried-only 0.2 wt.% Pt/TiO2 sample was

calcined in static air at 450 oC for 4 h (calc) and a second portion was calcined at 450

oC for 4 h and then reduced using 5%H2/Ar at 450 oC for 4 h (calc+red) and tested for

the hydrogenation of 3-NS under identical reaction conditions. The 0.2 wt.% Pt/TiO2

calc+red catalyst was found to be much more active (initial rate 6.04x10-3 mol3NS L-1

min-1) compared to the 0.2 wt.% Pt/TiO2 red-only catalyst. However, the selectivity to

3-VA for the calc+red catalyst was slightly lower (ca. 96%) compared to the red -

only catalyst, because of the generation of the fully hydrogenated product (3-

ethylaniline, 3-EA (4%)) (Figure S4). Corma et al.11 reported that a 0.2 wt.% Pt/TiO2

catalyst reduced at 450 oC displayed the growth of a TiO2 layer over the Pt nanoparticles resulting in a higher selectivity to 3-

VA (ca. 90%). However, this same material reduced at 200 °C did not show any evidence of TiOx over-growth on the Pt

nanoparticles and displayed a much lower selectivity for 3-VA (ca. 42%), with 3-EA (ca. 52%) being the main product.11 In

the present study, the 0.2% Pt/TiO2 red and 0.2% Pt/TiO2 calc+red materials, treated

at 450 °C, showed a difference in their catalytic activities only (Figure 1), but not in

their selectivities. In order to rationalise this observation, we prepared a series of

Pt/TiO2 catalysts with higher (0.5 wt.%) and lower (0.05 and 0.08 wt.%) Pt loadings.

All these catalysts underwent reduction-only (red) or calcination-only (calc) or

calcination followed by reduction (calc+red) heat treatments at 450 °C and were

tested for the hydrogenation of 3-NS. The initial rates for all these catalysts are

presented in Figure 1(a). For all the Pt loadings, the calc-only samples were found to

be completely inactive and hence, for clarity, the data for these catalysts are not

presented in Figures 1(a) & (b). For higher Pt loading catalysts (i.e., 0.2 and 0.5 wt.%)

the reaction rates of the red-only catalysts were found to be lower (1.55x10-3 mol3NS

L-1 min-1 and 2.84x10-3 mol3NS L-1 min-1 respectively) compared to their counterpart

calc+red catalysts (6.04x10-3 mol3NS L-1 min-1 and 1.11x10-2 mol3NS L-1 min-1

respectively). Interestingly, for the 0.05 wt.% Pt/TiO2 catalyst, the opposite trend was

observed whereby the red-only catalyst was found to be much more active (8.35x10-4

mol3NS L-1 min-1) compared to the calc+red catalyst (1.59x10-4 mol3NS L-1 min-1)

under identical reaction conditions. However, for the 0.08 wt.% Pt/TiO2 catalyst the

red and the calc+red samples displayed similar activity levels (i.e., 1.75x10-3 mol3NS

L-1 min-1 and 1.78x10-3 mol3NS L-1 min-1 respectively). It is important to highlight that

for all catalysts, irrespective of Pt loading or heat treatment protocol, the selectivity to

3-VA remained very high (> 90 %). It should also be noted that the catalysts with

higher Pt loadings (i.e., 0.2 and 0.5 wt.%) subjected to the calc+red treatment showed

slightly lower selectivity (because of the production of some 3-EA, which is an over

hydrogenation product) as compared to the corresponding red-only samples (see 3-

VA selectivity versus 3-NS conversion plot in Figure S4). These results demonstrate

that the post-synthesis heat treatment protocols combined with the Pt loading

modulate the active sites for this chemoselective hydrogenation reaction. The initial

rate data of all the red and calc+red Pt/TiO2 samples with different metal loadings are

reported in SI, Figure S1.

Figure 1: Effect of different heat treatments (i.e., red versus calc+red) on the activity

of Pt/TiO2 catalysts having different (i.e., 0.05, 0.08, 0.2, 0.5 wt.%) Pt loadings. (a)

Initial 3-NS conversion rates; (b) Turnover frequencies (TOF) based on the initial

rate obtained during the selective hydrogenation of 3-NS. Reaction conditions: 3-NS:

0.2 ml; toluene: 8ml; pH2: 3bar; catalyst: 50 mg; temperature: 40 °C.

Since the catalysts have different Pt loadings, their conversion rates were normalised

using the Pt loadings to allow for a meaningful comparison of their catalytic

activities. The resultant TOFs (Figure 1(b)) shows an exceptional intrinsic activity of

the 0.2 wt.% Pt/TiO2 calc+red catalyst (5650 mol3NS molPt-1 h-1). Comparing this TOF

value with other reported literature values (SI, Table S1) it is clear that our 0.2 wt.%

Pt/TiO2 calc+red material is the most active catalyst reported to date under

comparable reaction conditions. It should be noted that much higher TOFs have been

reported under harsher reaction conditions (i.e, higher temperatures, greater H2

pressure and higher metal loadings).29,30 In an effort to unravel the relationship

between metal loading, heat treatment protocol and the nanostructural properties of

these catalysts, all these materials were comprehensively characterised using a

complementary set of state-of-the-art microscopic and spectroscopic methods.

Electron microscopy studies

Figure 2 shows some representative HAADF-STEM images of red and calc+red

samples of Pt/TiO2 catalysts with different Pt loadings (i.e., 0.05, 0.08, 0.2 and 0.5

wt.%). The corresponding images of the calc series of samples are presented in SI,

Figure S5. This study allowed us to follow how the size and structure of the Pt

nanoparticles changes according to the heat treatment applied for each Pt loading

level. In all the calc-only samples, the vast majority of the Pt species are sub-nm

clusters and highly dispersed single Pt atoms distributed over the support (SI, Figure

S5). In the 0.05 wt.% Pt/TiO2 catalyst (Figures 2 (a) and (b)), the Pt particles found in

both the red-only and the calc+red were ultra-small clusters, however single Pt atoms

were not detected. The red-only sample had a more uniform particle size distribution

as compared to the calc+red catalyst, which correlates well with the observed

catalytic activity trend. In the case of the 0.08 wt.% Pt/TiO2 material (Figures 2(c) and

(d)), both red-only and calc+red samples possess a very similar nanostructure with a

uniformly small mean Pt particle size (1-2 nm) and reasonably similar particle size

distributions. This observation again correlates well with the very similar catalytic

activities noted for the red-only and calc+red samples of the 0.08 wt.% Pt/TiO2

catalyst (Figure 1). In the case of 0.2 wt.% Pt/TiO2 catalyst (Figures 2(e) and (f)), the

calc+red sample showed many uniformly distributed small Pt particles (1-2 nm)

distributed over the TiO2 support. However, the corresponding red-only sample

showed a wider distribution of Pt particle sizes (1-4 nm) with a larger mean particle

diameter (1.6 nm). In the case of the 0.5 wt.% Pt/TiO2 catalyst (Figures 2(g) and

2(h)), the red-only sample showed a mixture of smaller particles (ca. 1-2 nm) and

larger particles (>5 nm), whereas in the corresponding calc+red variant, uniform

particles are homogeneously distributed over the TiO2 surface. The possibility of

improving the metal dispersion in favour of small Pt particles by performing

consecutive calcination and reduction treatments has been observed before.26,27 Our

microstructural observations correlate very well with the measured catalytic activites

for these various Pt/TiO2 catalysts, indicating that smaller Pt particles with high

dispersion and narrow size distribution lead to a higher overall catalytic activity.

Analysis of the particle size distribution data for all the samples in this study (Figure

2) reveals that those Pt/TiO2 catalysts having the largest population of ~1 nm particles

and narrowest size distribution (i.e., 0.08 wt.% Pt/TiO2 (red), 0.08 wt.% Pt/TiO2

(calc+red) and 0.2 wt.% Pt/TiO2 (calc+red)) also showed the highest catalytic

activity in terms of TOF values (Figure 1(b)).

Figure 2: Representative HAADF-STEM images and the derived particle size

distributions of the unused Pt/TiO2 catalysts; (a) 0.05 wt.% Pt/TiO2 (calc+red), (b)

0.05 wt.% Pt/TiO2 (red), (c) 0.08 wt.% Pt/TiO2 (calc+red), (d) 0.08 wt.% Pt/TiO2

(red), (e) 0.2 wt.% Pt/TiO2 (calc+red), (f) 0.2 wt.% Pt/TiO2 (red), (g) 0.5 wt.%

Pt/TiO2 (calc+red) and (h) 0.5 wt.% Pt/TiO2 (red).

X-ray photoelectron spectroscopy studies

All the Pt/TiO2 catalysts with different loadings and heat treatment protocols were

characterised by XPS and the results are shown in Figure 3. It should be noted that the

samples with very low loading levels (i.e., < 0.1wt.%) , gave noisy spectra even after

an unusually large number of scans (>100). This signal-to-noise ratio problem was

compounded by the presence of an overlapping Ti loss peak at ca. 75 eV which made

quantification of these particular spectra quite challenging.28

Generally, regardless of Pt loading, the XPS data were quite similar, therefore for

conciseness of discussion we will focus initially on the data acquitred from the 0.5

wt.% Pt/TiO2 catalysts. For the dried-only sample, a Pt(4f7/2) binding energy of 72.2

eV was measured, which is indicative of a Pt(II) species, presumably associated with

residual chlorine as detected by XPS, although the presence of Pt oxides cannot be

excluded.29,30 Calcination of this highest loading sample revealed two distinct Pt

environments; the first with a binding energy again of 72.2 eV and a second with a

binding energy of 74.4 eV. Considering the calcination temperature employed (450

°C), the small concentration of Cl detected (ca. 0.2 – 0.25 at.%, see SI, Table S2), and

the nominal Pt concentration (ca. 0.2 at.%, see SI, Table S2) a bulk PtCl2 species is

not the cause of the 72.2 eV signal, therefore this is likely a PtOx phase. The higher

binding energy component at 74.4 eV, can be assigned as PtO2 or Pt hydroxide.31

Both ‘red’ and ‘calc + red’ catalysts exhibit Pt(0), as indicated by the asymmetric

peak shape and a binding energy of 70.6 – 70.8 eV.32 Representative HAADF-STEM

lattice images of the 0.5 wt.% Pt/TiO2 (red), (presented in SI, Figure S6), confirms the

presence of metallic Pt(0) structure in larger particles.

Figure 3: XPS data of the (a) dried only, (b), calcined-only, (c) reduced-only and (d)

calcined+reduced Pt/TiO2 catalysts with different Pt loading levels (i.e., (i) 0.05 wt.%; (ii)

0.08 wt.%; (iii)0.2 wt.%; (iv) 0.5 wt.%).

For the low (0.08 and 0.05 wt.%) Pt loading catalysts, depending on the heat

treatments, the Pt(4f) signal strength was very weak, especially for those materials

that had undergone reductive treatments, which is attributed to changes in the particle

size.

X-ray absorption spectroscopy studies

XAFS measurements were performed for the 0.05, 0.08, 0.2 and 0.5 wt.% Pt/TiO2

(red & calc+red) samples. Analysis of the XAFS data included interpretation of both

XANES and EXAFS information. The platinum L3 edge XANES spectrum is

dominated by the dipole allowed 2p3/2 5d5/2 transition of a core electron, which is

known to be sensitive to the oxidation and co-ordination state of Pt.33-36 In Figures 4

(a) and (b), the XANES plots show a comparison of samples with different metal

loadings that underwent the same heat treatment: namely red-only (Figure 4(a)) or

calc+red (Figure 4(b)). Figure S7 also contains a plot of the XANES for the 0.05

wt.% Pt/TiO2 samples and the PtO2 reference material. Already from this overview it

is possible to notice the variety of platinum species and environments present in these

samples.

11550 11555 11560 11565 11570 11575 115800.0

0.5

1.0

1.5

2.0

Nor

mal

ised

inte

nsity

Energy (eV)

0.5 % Pt 0.2 % 0.08 % 0.05 %

a

11550 11555 11560 11565 11570 11575 115800.0

0.5

1.0

1.5

2.0 0.5 % Pt 0.2 % 0.08 % 0.05 %

Nor

mal

ised

Inte

nsity

Energy (eV)

b

Figure 4. Pt L3-edge XANES (a and b) and k3 EXAFS FT (c and d) data recorded on Pt/TiO2

samples after reduction only (a, c) and calcination + reduction (b, d). Note: dashed arrow

indicates the decrease in edge position with increased loading. The solid arrow indicates the

increased rising absorption edge. Labels above the peaks indicate the scattering pairs that

give rise to that contribution.

The XANES spectra in Figure 4(a) appear almost identical for the three lowest Pt

loadings (i.e., 0.05, 0.08 and 0.2 wt.%) with a maximum in the µ(E) trace seen at ~

11,568 eV. It is noticeable however that the maximum of the rising absorption edge of

the samples and the edge position is lower (i.e., ‘red’-shifted) compared that of the

reference PtO2 phase (Figure S7). The sample with 0.5 wt.% Pt loading is especially

red-shifted and its rising absorption edge possesses the lowest overall intensity,

particularly at 11,568 eV. Previously it has been shown that such changes indicate

differences in the extent of reduction/oxidation with all samples possessing an

average Pt oxidation state < + 4 with the 0.5 wt.% sample being more reduced than all

samples with lower loadings, which are seen to be the most oxidised.37 These

1 2 3 4 5 60

2

4

6

8

10

Pt-O

0.5 % Pt 0.2 % 0.08 % 0.05 %

FT M

agni

tude

(k3 (

).k)

R (Å)

Pt-Ptc

1 2 3 4 5 60

2

4

6

8

10

0.5 % Pt 0.2 % 0.08 % 0.05 %

FT M

agni

tude

(k3 (

).k)

R (Å)

Pt-O

d

observations are entirely consistent with the XPS and HAADF-STEM data shown in

Figures 3 and S6 respectively.

A similar situation can be seen for the calc+red samples shown in Figure 4(b),

although, based on the intensity of the rising absorption edge, all samples except that

containing 0.08 wt.% Pt appear more oxidised than after simple reduction; the 0.08

wt.% sample appearing to remain the same. In order to verify and even extract further

information from these samples, the EXAFS Fourier Transform (FT) data are plotted

in Figures 4 (c) and (d). Focusing on Figure 4(c) of the reduced-only samples, it is

clear that all except the 0.5 wt.% Pt/TiO2 red sample contain a strong peak at ~ 2.0 Å

assignable to a Pt-O distance typical of PtO2 species; see Figure S8(a) which

illustrates the similarities of the Pt-O distances in these samples compared to that of

PtO2.38,39 The intensity of the Pt-O contribution with respect to the reference PtO2 is

much smaller in all the reduced samples and suggests that the Pt-O environments in

the samples are more disordered. This can be rationalised as due to static disorder

caused by the high dispersion of PtO2 species across the TiO2 support and according

to the XANES, a partial reduction of the Pt4+ to Pt2+. Note that since a PtO reference is

not readily available, it is not possible to unambiguously identify its presence nor rule

out its absence. The 0.5 wt.% Pt/TiO2 red-only sample also contains a broad peak

with a maximum intensity at ~ 2.73 Å typical of Pt metal. Analysis of the EXAFS

data (see Figure S9 and Table S3) for this sample suggests the existence of Pt

nanoparticles ~ 1 nm in diameter.40 For the 0.05 and 0.2 wt.% Pt/TiO2 red-only

samples, weak peaks at 2.87 and 2.81 Å may also suggest the presence of a small

amount of Pt metal in these samples, as well as a much more significant Pt-oxide

environment.

Figure 4(d) contains FT data of the corresponding calc+red samples and are

characterised by the presence of a strong Pt-O contribution in all samples, although

the intensity of this contribution is inversely proportional to the amount of Pt present;

see Figure S8(b) which illustrates the similarities of the Pt-O distances in the samples

to that of the reference PtO2 material. This is again consistent with the XANES data,

which suggests the presence of more Pt4+ in the samples after the calc+red treatment;

the 0.05 wt.% Pt/TiO2 calc+red sample in particular contains Pt-Pt scattering

contributions consistent with PtO2 suggesting that in this sample some PtO2

nanocrystallites may have evolved. For the 0.5 wt% Pt/TiO2 calc+red sample, the Pt-

O peak possesses a skew and kurtosis towards higher Pt-O distances and when

considered along with the weaker absorption edge intensity for this sample, suggests

the presence of more Pt2+. Interestingly no real evidence for metallic Pt species could

be seen on visual inspection or when attempts were made to fit the data.

CO DRIFTS analysis

CO DRIFTS analysis was performed on all of the Pt/TiO2 catalysts to gain an accurate

measure of the degree of Pt metal exposure on the surface of the catalysts. For all the

samples, a broad asymmetric band was observed between 2050 – 2060 cm-1 and for

all the calc+red samples, a sharp band at ca. 2089 cm-1 was also present (Figure 5). It

has been reported that the band at ca. 2050-2060 cm-1 corresponds to CO linearly

adsorbed on low coordination Pt edge and corner sites.11,41-44 This IR band is quite

sensitive to a number of factors including Pt dispersion, Pt microstructure (i.e., steps,

crystal face, etc.) and charge transfer in the vicinity of the Pt adsorption site. The band

at ca. 2089 cm-1 corresponds to CO molecules adsorbed on Pt (111) terrace sites

having a coordination number of 9.11,43-45 The Pt/TiO2 system has also been widely

reported to display a strong metal support interaction (SMSI), especially when

reduced at temperatures higher than 400 °C, which typically results in a much lower

uptake or adsorption of CO.46 The 0.5 wt.% Pt/TiO2 red-only and 0.2 wt.% Pt/TiO2

red-only samples exhibit substantially less CO adsorption confirming the presence of

an SMSI effect in these samples. The SMSI could be associated with the formation of

bonds between Pt metal and cationic or atomic Ti species or by coverage of the

platinum particle by a layer of TiOx, as a result of hydrogen spill-over. Corma et al.

has reported TiO2 coverage over Pt particles for a 0.2 wt.% Pt/TiO2 sample reduced at

450 °C.11 Combining our CO DRIFTS data with reported interpretations, we consider

that the Pt nanoparticles are covered by TiOx for the 0.5 wt.% Pt/TiO2 red-only and

0.2 wt.% Pt/TiO2 red-only samples, even though this layer was not resolved by

STEM. Analysing the TOF values of these two samples from Figure 1, it is clear that

their activities are much lower than all of the other samples tested (1060 mol3NS molPt-

1 h-1 and 1450 mol3NS molPt-1 h-1, respectively). Although some SMSI effect can be

beneficial for increasing chemoselectivity in this reaction, an excessive TiOx

overlayer on the Pt can be detrimental to the overall catalytic activity through

complete site blocking. For the 0.5 wt.% Pt/TiO2 and 0.2 wt.% Pt/TiO2 catalysts

subjected to the calc+red treatment, the Pt particles were not completely covered by

TiOx leaving some Pt sites exposed for catalytic turnover. Although we could not

directly visualise the surface coverage of Pt nanoparticles by TiOx, CO DRIFTS data

provides convincing evidence for some TiOx coverage of Pt particles in the 0.5 wt.%

Pt/TiO2 and 0.2 wt.% Pt/TiO2 red-only samples.

Figure 5. CO DRIFTS spectra of the (i) 0.05 wt.%, (ii), 0.08 wt.%, (iii), 0.2 wt.% and (iv) 0.5

wt.% Pt/TiO2 catalysts in (a) the reduced only and (b) calcined + reduced states.

It has previously been reported that peripheral sites at the Pt/TiO2 interface are

important for some selective hydrogenation reactions such as the hydrogenation of

crotonaldehyde and furfuraldehyde.47,48 Following this observation, the total number

of peripheral sites for all the Pt/TiO2 catalysts examined were estimated by assuming

all Pt particles are half-‘sphere’ Mackay icosahedra (see SI, Table S4), based on the

particle size distributions derived from HAADF-STEM images. The total number of

Pt peripheral sites, expressed as a weight percentage of the total catalyst mass, for the

various Pt/TiO2 materials, are plotted in Figure 6(a) together with the 3-NS initial

conversion rate.

2150 2100 2050 2000 19500.00

0.02

0.04

0.06

0.08

0.10

2086

2090

2053

a

ii) 0.08wt% Pt/TiO2

Inte

nsity

(a.u

.)

Wavenumber (cm-1)

b

2050

2150 2100 2050 2000 19500.00

0.01

0.02

0.03

0.04

0.05

a

Inte

nsity

(a.u

)

Wavenumber (cm-1)

b

i) 0.05wt% Pt/TiO2

2087

20602042

2051

2085

2150 2100 2050 2000 19500.00

0.05

0.10

0.15iii) 0.2wt% Pt/TiO2

b

Inte

nsity

(a.u

.)

Wavenumber (cm-1)

a

20522060

2089

2053

2082

2150 2100 2050 2000 19500.00

0.05

0.10

0.15

0.20

0.25

0.30

2056

a

iv) 0.5wt% Pt/TiO2

Inte

nsity

(a.u

.)

Wavenumber (cm-1)

b

20502060

2089

2104

Figure 6. The correlation between the 3-NS initial conversion rate and total amount of Pt

present in (a) peripheral sites and (b) as exposed surface atoms for the different Pt/TiO2

catalysts as estimated from analysis of HAADF-STEM images. For model (a) R2 = 0.89,

p<0.001, and for model (b) R2 = 0.71, p<0.02.

A highly significant positive correlation is observed between the number of peripheral

Pt sites and catalytic activity (Figure 6(a)), supporting the proposition of Boronat et

al. that the peripheral sites created where the Pt nanoparticles make contact with the

TiO2 support are very important for the selective hydrogenation of nitrogroups.49 A

weaker fit is observed when the same initial rate data was plotted against the total

number of exposed surface Pt atoms (Figure 6(b)). The Pt surface area determined

through CO chemisorption was plotted against the initial rate (SI, Figure S11) also

gave a weaker correlation. These data from two different methods indicate that the Pt

surface atoms may not be the active sites. Post synthetic heat treatment plays an

important role in inducing SMSI effects in Pt/TiO2 catalysts, which in turn has a

profound impact on their resultant catalytic activities. High temperature reduction

induces coverage of Pt sites by TiOx. This problem can be eliminated however by

employing a sequential calcination step followed by reduction step. We propose that

the calcination step prior to the reduction treatment more effectively anchors the Pt

particles to the support and thus limits their subsequent mobility during the reduction

treatment. From the particle size distribution data (Figure 2), it is evident that, among

all the catalysts tested, only 0.5wt%Pt/TiO2 – red & 0.2 wt.% Pt/TiO2 – red samples

have metal particles larger than 2-3 nm. From CO-DRIFTS data (Figure 5), only these

two samples show signs of TiOx coverage over Pt particles. This suggest a possible

size dependant TiOx coverage phenomenon which needs further detailed

investigation. With the current data we conclude that for samples with higher Pt

loadings, a calcination+reduction treatment results in more active catalysts compared

to reduced-only catalysts. In the case of lower Pt loadings, due to high dispersion and

relative scarcity of nearby Pt neighbours, even with a reduction-only treatment, there

is little agglomeration of the Pt nanoparticles in spite of Pt mobility. Hence, the

calcination pre-step is not necessary for samples with low Pt loadings. The collective

electron microscopy and spectroscopic data obtained from all the Pt/TiO2 samples

suggests that those materials with very small Pt particles (~ 1 nm) and without any

significant TiOx surface coverage display exceptional catalytic performance.

Conclusions

Pt nanoparticles supported on TiO2 with different Pt loadings, were found to be

remarkably active and selective for the chemoselective hydrogenation of 3-NS to 3-

VA. Amongst all the catalysts tested, the 0.2 wt.% Pt/TiO2 calc+red material was

found to be the most active. It was also the catalyst with the highest number of

peripheral active sites generated as a result of the combination of ultra-small Pt

nanoparticles (~1 nm) with a tight particle size distribution in which the the Pt

particles were not covered by TiOx. This nanostructure was achieved by combining a

calcination treatment at 450 °C with a subsequent reduction of this catalyst at 450 °C,

thereby preventing any TiOx coverage of the Pt active sites. For the same 0.2 wt.%

level of Pt loading, a reduction treatment alone at 450 °C induces an SMSI effect

which results in the partial coverage of active Pt sites and a less active catalyst.

Methods

Catalyst Preparation

Supported Pt catalysts with different Pt loadings (0.05, 0.08, 0.2 and 0.5 wt.%) were

synthesised using a wet-impregnation method. The dried materials were then heat

treated in a number of ways and tested for the chemoselective hydrogenation of 3-

nitrostyrene. For a typical catalyst preparation (2g), the requisite amount of aqueous

hydrogen hexachloroplatinic acid solution (9.6 mg Pt/ml, assay 30.21% H2PtCl6,

Johnson Matthey) was diluted to 16 ml using deionised water and heated in an oil

bath to 60 °C with vigorous stirring (800 rpm) before addition of the requisite amount

of support (TiO2, Evonik®P25). The temperature was then increased to 95 °C for 16 h

to dry the slurry in air. The resultant material was then ground thoroughly (designated

as ‘fresh dried-only’ sample). A portion (500 mg) of the dried-only sample was either

calcined in static air or reduced under a steady flow of 5% H2/Ar. Heat treatments

were carried out at 450 °C for 4 h with a heating rate of 10 °C min-1. Samples that

underwent reduction-only were denoted as “red”, calcination only as “calc” and

calcination followed by reduction treatment as “calc+red”. All these materials were

tested as catalysts without any further treatment unless stated otherwise.

Hydrogenation of 3-nitrostyrene

Liquid phase chemoselective hydrogenation of 3-nitrostyrene was performed in a 50

ml glass round bottom Colaver® reactor. In a typical reaction, 3-nitrostyrene (0.2 ml),

toluene (8 ml) and catalyst (50 mg) were charged in the reactor which was then

purged with N2 to remove traces of air and immersed in an oil bath set at 40 °C. After

purging the reactor three times with 3 bar H2 while stirring, the reaction was

commenced. After a pre-determined reaction time, the reactor was cooled to <5 °C

using ice, after which aliquots of the reaction mixture were collected and centrifuged

to remove the solid catalyst. The liquid reaction mixture (1 ml) was added to a GC

vial together with a fixed amount of external standard o-xylene (0.1 ml) for further

quantitative analysis. The determination of the conversion was performed using a

Varian 450GC Gas Chromatograph equipped with a HP-5ms boiling point column, a

Flame Ionisation detector (FID) and a CP8400 Autosampler. Conversion of the

substrate and product selectivity were calculated using suitable calibration plots and

response factors.

Catalyst Characterization

X-ray Photoelectron Spectroscopy (XPS)

XPS was performed using two systems; the first was a Kratos Axis Ultra DLD

spectrometer using monochromatic Al K radiation (source power 140 W). An

analyser pass energy of 160 eV was used for survey scans, and 40 eV for detailed

acquisition of individual elemental regions. The second system was a Thermo

Scientific K-Alpha+ spectrometer utilizing microsfocused monochromatic Al K

radiation (source power 72 W), with analyser pass energies of 40 and 150 eV. Data

was analysed with CasaXPS, using escape depth corrected sensitivity factors supplied

by the instrument manufacturers.

Scanning Transmission Electron Microscopy

Samples for examination by STEM were prepared by dry dispersion of the catalyst

powder onto a holey carbon film supported by a 300 mesh copper TEM grid. Bright

field (BF) and high angle annular dark field (HAADF) STEM images were taken

using an aberration corrected JEM ARM-200CF microscope operating at 200kV. This

instrument was also equipped with a JEOL Centurio silicon drift detector for X-ray

energy dispersive spectroscopy (XEDS). Particle size distribution analysis was

performed from analysis of HAADF electron micrographs using Image J.

X-ray Absorption Spectroscopy

XAFS spectroscopy measurements were performed on stations B18 and I20 at the

Diamond Light Source (Harwell, Oxon) operating at 3 GeV in multibunch mode with

a current of 200 mA. Both stations are equipped with a Si(111) double crystal

monochromator, ion chambers for measuring incident and transmitted beam

intensities and/or a multi-element Ge (fluorescence) detector for recording X-ray

absorption spectra. The measurements were carried out in air on self-supporting

wafers (approximately 100 mg) at the Pt L3 edge, and a 10 µm thick Pt foil was used

as a calibration sample for the monochromator. Measurements were performed at

room temperature in normal step scanning mode. To improve the signal-to-noise ratio,

multiple scans were taken. All data were subjected to background correction using

Athena (i.e., IFEFFIT software package for pre- and post-edge background

subtraction and data normalisation). XAFS spectra were normalized from 30 to 150

eV above the edge energy, while the EXAFS spectra were normalized from 150 eV to

the last data point using the Autobk algorithm. Normalisation was performed between

μ(E) and μ0(E) via a line regression through the data in the region below the edge and

subtracted from the data. A quadratic polynomial was then regressed to the data above

the edge and extrapolated back to E0. The extrapolated value of the post edge

polynomial at E0 was used as the normalisation constant. This threshold energy (E0) is

normally determined using either the maximum in the 1st derivative, approximately 50

% of the rising absorption edge, or immediately after any pre-edge or shoulder

features. The isolated EXAFS spectra were analysed using the DL-EXCURV

programme. Data was analysed using a least squares single or dual shell EXAFS

fitting analysis performed on data that had been phase corrected using muffin-tin

potentials. Amplitude reduction factors (So2) of 0.94 obtained from fitting a Pt metal

foil, was also used in the analysis.

CO DRIFTS

For CO-DRIFTS analysis, samples were loaded in a DRIFT Tensor 27 spectrometer,

and flushed with N2 for 30 min to record a baseline. The sample was then exposed to

20% CO/Ar until complete saturation of the Pt sites was achieved. The system was

then flushed with pure N2 until there was no presence of atmospheric CO and then the

spectra were recorded.

Acknowledgements:

MM and AJB acknowledge MAXNET Energy consortium for funding. AJB also

acknowledges the EPRSC Centre for Doctoral Training in Catalysis (Grant number:

EP/L016443/1). MS and QH thank Cardiff University for their respective University

Research Fellowships and RQ acknowledges Chinese Scholarship Council (CSC)

funding for his stay at Cardiff University. CJK gratefully acknowledges funding from

the National Science Foundation Major Research Instrumentation program (GR#

MRI/DMR-1040229). SMA thanks the Saudi Arabian government for his PhD

scholarship. The authors thank the UK Catalysis Hub for allocating beamtime slots

through the BAG allocation for X-ray acquisition of the absorption spectroscopic data

at the Diamond synchrotron facility. We are also indebted to Dr. Peter Wells for

acquisition of the X-ray absorption spectroscopic data at the Diamond synchrotron

facility. The authors thank the Diamond Light Source for access to the electron

Physical Science Imaging Centre (ePSIC Instrument E01 and proposal number

MG22776) that contributed to the results presented here.

Authors contribution:

MM, AJB and RQ performed the catalyst preparation and catalyst testing under the

supervision of MS, ND, SF and XG. SMA, QH and CJK carried out electron

microscopy studies of the catalyst. DJM performed the XPS characterisation of the

catalysts. EG and AMB analysed the XAS data. DB suggested the kinetics

experiments presented in this article. MS conceived this idea and supervised the

project. GJH directed this project. All authors contributed to the data analysis and

drafting of this manuscript.

Competing interests

The authors declare no competing financial interests.

References

1 Wang, L. et al. Single-site catalyst promoters accelerate metal-catalyzed

nitroarene hydrogenation. Nature Communications 9, 1362, (2018).

2 Therrien, A. J. et al. An atomic-scale view of single-site Pt catalysis for low-

temperature CO oxidation. Nature Catalysis 1, 192-198, (2018).

3 Vile, G., Albani, D., Almora-Barrios, N., Lopez, N. & Perez-Ramirez, J.

Advances in the design of nanostructured catalysts for selective

hydrogenation. Chem.Cat.Chem. 8, 21-33, (2016).

4 Blaser, H. U., Steiner, H. & Studer, M. Selective catalytic hydrogenation of

functionalized nitroarenes: An update. Chem.Cat.Chem. 1, 210-221, (2009).

5 Song, J. J. et al. Review on selective hydrogenation of nitroarene by catalytic,

photocatalytic and electrocatalytic reactions. Applied Catalysis B -

Environmental 227, 386-408, (2018).

6 Rylander, P. N. Hydrogenation Methods. (Academic Press, 1990).

7 Nishimura, S. Handbook of Heterogeneous Catalytic Hydrogenation for

Organic Synthesis (Wiley-VCH, 2001).

8 Bellamy, F. D. & Ou, K. Selective reduction of aromatic nitro-compounds

with stannous chloride in non-acidic and non-aqueous medium. Tetrahedron

Letters 25, 839-842, (1984).

9 Basu, M. K., Becker, F. F. & Banik, B. K. Ultrasound-promoted highly

efficient reduction of aromatic nitro compounds to the aromatic amines by

samarium/ammonium chloride. Tetrahedron Letters 41, 5603-5606, (2000).

10 Burawoy, A. & Critchley, J. P. Electronic spectra of organic molecules and

their interpretation-V*. Tetrahedron 5, 340-351, (1959).

11 Corma, A., Serna, P., Concepcion, P. & Calvino, J. J. Transforming non-

selective into chemoselective metal catalysts for the hydrogenation of

substituted nitroaromatics. Journal of the American Chemical Society 130,

8748-8753, (2008).

12 Arnold, H., Döbert, F. & Gaube, J. in Handbook of Heterogeneous Catalysis

(Wiley-VCH, 2008).

13 Shi, Q. X., Lu, R. W., Jin, K., Zhang, Z. X. & Zhao, D. F. Simple and eco-

friendly reduction of nitroarenes to the corresponding aromatic amines using

polymer-supported hydrazine hydrate over iron oxide hydroxide catalyst.

Green Chemistry 8, 868-870, (2006).

14 Shi, Q., Lu, R., Lu, L., Fu, X. & Zhao, D. Efficient reduction of nitroarenes

over nickel-iron mixed oxide catalyst prepared from a nickel-iron hydrotalcite

precursor. Advanced Synthesis & Catalysis 349, 1877-1881, (2007).

15 Wang, X. K. et al. Heterogeneous catalytic transfer partial-hydrogenation with

formic acid as hydrogen source over the schiff-base modified gold nano-

catalyst. Catalysis Letters 147, 517-524, (2017).

16 Pogorelic, I. et al. Rapid, efficient and selective reduction of aromatic nitro

compounds with sodium borohydride and Raney nickel. J. Mol. Catal. A -

Chem. 274, 202-207, (2007).

17 Rahman, A. & Jonnalagadda, S. B. Swift and selective reduction of

nitroaromatics to aromatic amines with Ni–boride–silica catalysts system at

low temperature. Catalysis Letters 123, 264-268, (2008).

18 Corma, A. & Serna, P. Chemoselective hydrogenation of nitro compounds

with supported gold catalysts. Science 313, 332-334, (2006).

19 Wei, H. et al. FeOx-supported platinum single-atom and pseudo-single-atom

catalysts for chemoselective hydrogenation of functionalized nitroarenes.

Nature Communications 5, 5634 - 5642, (2014).

20 Wei, H. et al. Remarkable effect of alkalis on the chemoselective

hydrogenation of functionalized nitroarenes over high-loading Pt/FeOx

catalysts. Chemical Science 8, 5126-5131, (2017).

21 Sankar, M. et al. Synthesis of stable ligand-free gold-palladium nanoparticles

using a simple excess anion method. ACS Nano 6, 6600-6613, (2012).

22 Zanella, R., Louis, C., Giorgio, S. & Touroude, R. Crotonaldehyde

hydrogenation by gold supported on TiO2: structure sensitivity and

mechanism. Journal of Catalysis 223, 328-339, (2004).

23 Cargnello, M. et al. Efficient removal of organic ligands from supported

nanocrystals by fast thermal annealing enables catalytic studies on well-

defined active phases. Journal of the American Chemical Society 137, 6906-

6911, (2015).

24 Freakley, S. J. et al. Palladium-tin catalysts for the direct synthesis of H2O2

with high selectivity. Science 351, 965-968, (2016).

25 Chen, C., Cao, J. J., Cargnello, M., Fornasiero, P. & Gorte, R. J. High-

temperature calcination improves the catalytic properties of alumina-supported

Pd@ceria prepared by self assembly. Journal of Catalysis 306, 109-115,

(2013).

26 Matos, J. et al. In situ coarsening study of inverse micelle-prepared Pt

nanoparticles supported on gamma-Al2O3: pre-treatment and environmental

effects. Physical Chemistry Chemical Physics 14, 11457-11467, (2012).

27 Kim, G. J., Kwon, D. W. & Hong, S. C. Effect of Pt particle size and valence

state on the performance of Pt/TiO2 catalysts for CO oxidation at room

temperature. Journal of Physical Chemistry C 120, 17996-18004, (2016).

28 Zimmermann, R. et al. Electronic structure of 3d-transition-metal oxides: On-

site Coulomb repulsion versus covalency. Journal of Physics-Condensed

Matter 11, 1657-1682, (1999).

29 Riggs, W. M. X-ray photoelectron spectrometry of platinum compounds.

Analytical Chemistry 44, 830-832, (1972).

30 Moulder, J. F., Stickle, W. F., Sobol, P. E. & Bomben, K. D. Handbook of X-

ray Photoelectron Spectroscopy. (Perkin-Elmer Corporation, 1992).

31 Naumkin, A. V., Kraut-Vass, A., Gaarenstroom, S. W. & Powell, C. J.

(National Institute of Standards and Technology, Gaithersburg MD, 2000).

32 Janin, E. et al. Hydrogen adsorption on the Pt(111)( 3 × 3 ) R30° Sn surface

alloy studied by high resolution core level photoelectron spectroscopy.

Applied Surface Science 99, 371-378, (1996).

33 Brown, M., Peierls, R. E. & Stern, E. A. White lines in X-ray absorption.

Physical Review B 15, 738-744, (1977).

34 Koningsberger, D. & Prins, R. X-ray absorption: Principles, applications,

techniques of EXAFS, SEXAFS, and XANES. (Wiley, 1988).

35 Penner-Hahn, J. E. X-ray absorption spectroscopy in coordination chemistry.

Coordination Chemistry Reviews 190, 1101-1123, (1999).

36 de Groot, F. High-resolution X-ray emission and X-ray absorption

spectroscopy. Chemical Reviews 101, 1779-1808, (2001).

37 Yoshida, H., Nonoyama, S., Yazawa, Y. & Hattori, T. Quantitative

determination of platinum oxidation state by XANES analysis. Physica

Scripta T115, 813-815, (2005).

38 Hyde, T. I. et al. X-ray absorption spectroscopic studies of platinum speciation

in fresh and road aged light-duty diesel vehicle emission control catalysts.

Platinum Metals Review 55, 233-245, (2011).

39 Kim, M. Y., You, Y. S., Han, H. S. & Seo, G. Preparation of highly dispersed

platinum catalysts impregnated on titania-incorporated silica support.

Catalysis Letters 120, 40-47, (2008).

40 Beale, A. M. & Weckhuysen, B. M. EXAFS as a tool to interrogate the size

and shape of mono and bimetallic catalyst nanoparticles. Physical Chemistry

Chemical Physics 12, 5562-5574, (2010).

41 Raskó, J. CO-induced surface structural changes of Pt on oxide-supported Pt

catalysts studied by DRIFTS. Journal of Catalysis 217, 478-486, (2003).

42 Zafeiratos, S., Papakonstantinou, G., Jacksic, M. M. & Neophytides, S. G. The

effect of Mo oxides and TiO2 support on the chemisorption features of linearly

adsorbed CO on Pt crystallites: an infrared and photoelectron spectroscopy

study. Journal of Catalysis 232, 127-136, (2005).

43 DeRita, L. et al. Catalyst architecture for stable single atom dispersion enables

site-specific spectroscopic and reactivity measurements of CO adsorbed to Pt

atoms, oxidized Pt clusters, and metallic Pt clusters on TiO2. Journal of the

American Chemical Society 139, 14150-14165, (2017).

44 Thang, H. V., Pacchioni, G., DeRita, L. & Christopher, P. Nature of stable

single atom Pt catalysts dispersed on anatase TiO2. Journal of Catalysis 367,

104-114, (2018).

45 Stakheev, A. Y., Shpiro, E. S., Tkachenko, O. P., Jaeger, N. I. & Schulz-

Ekloff, G. Evidence for monatomic platinum species in H-ZSM-5 from FTIR

spectroscopy of chemisorbed CO. Journal of Catalysis 169, 382-388, (1997).

46 S. J. Tauster, S. C. F., and R. L. Garten. Strong metal-support interactions.

Group 8 noble metals supported on TiO2. Journal of the American Chemical

Society 100, 170-175, (1978).

47 Vannice, M. A. & Sen, B. Metal support effects on the intramolecular

selectivity of crotonaldehyde hydrogenation over platinum. Journal of

Catalysis 115, 65-78, (1989).

48 Baker, L. R. et al. Furfuraldehyde hydrogenation on titanium oxide-supported

platinum nanoparticles studied by sum frequency generation vibrational

Ssectroscopy: Acid–base catalysis explains the molecular origin of strong

metal–support interactions. Journal of the American Chemical Society 134,

14208-14216, (2012).

49 Boronat, M. & Corma, A. Origin of the different activity and selectivity

toward hydrogenation of single metal Au and Pt on TiO2 and bimetallic Au-

Pt/TiO2 catalysts. Langmuir 26, 16607-16614, (2010).